Vous êtes sur la page 1sur 5


The Whitmore Section

By William A. Thornton, P.E., Ph.D., and Carlo Lini, P.E.

How to use the Whitmore method for tension and compression strength checks.

Anyone who has had the task of designing a

bracing or truss connection has probably come across the
Whitmore section. For most cases, the method is simple and
straightforward. However, there are situations where deter-
mining the Whitmore section along with the tension and/or
compression checks that follow are not quite so clear. This
article addresses potential areas of confusion, and provides
Fig. 1: Truss outline (Whitmore, 1952).

➤ ➤
the reader with background information on the develop-
ment of this approach. Fig. 2: Stress distribution using Whitmore method (Whitmore, 1952).

Whitmore 101
First, it’s important to recognize what the Whitmore sec-
tion is. It is a simple way to determine how force from a brace
spreads through a gusset plate. It’s used to make checks of
gusset plate yielding and buckling possible.
Gusset plates have been used in steel structures since the
earliest metal trusses. However, research in the early part
of the 20th century regarding the distribution of stresses in
gusset plates under tension or compression loading was lim-
ited. R.E. Whitmore made note of this lack of knowledge in
1952, providing the following 1941 quote from T.H. Rust,
who had conducted earlier tests on gusset plates:
“It is difficult to believe that there is a more impor-
tant or more fundamental problem in need of further
investigation in the field of structural engineering
than steel gusset plates. They constitute a formidable
problem in stress analysis capable of further exploita-
tion in the laboratory…”
In an attempt to better understand gusset plates, Whitmore Department at the University of Tennessee during that time,
conducted a test on a mock-up of a truss joint connection had been interested in gusset plate stresses, most likely as a
for a 295-ft truss that was constructed at quarter scale (see result of his work experience in bridge design while at Ash,
Figure 1). Armour T. Granger, head of the Civil Engineering Howard-Needles and Tammen. It was upon Granger’s sug-
gestion, and under his supervision, that
Whitmore conducted this test. Based on
the test results, Whitmore concluded that
William A. Thornton, P.E., stresses occurred on the gusset plate as
Ph.D., is corporate consultant to shown in Figure 2.
AISC member firm Cives Engi- This article focuses on the direct
neering Corporation, Roswell, tension and compression stresses on
Ga. Carlo Lini, P.E., is an the Whitmore section. It is also impor-
AISC Steel Solutions Center tant to note that some of the conditions
advisor. Prior to joining AISC presented in this article were not tested
in 2011 he was a staff engineer by Whitmore, but are what we believe
with Ruby + Associates, Farm- are reasonable answers to questions we
ington Hills, Mich. have received.


Although Whitmore’s findings were published in May, 1952, Example II.C-2 in the AISC Design Examples illustrates the pro-
widespread use of the Whitmore section did not occur until the cess. Additionally, Examples II.C-1, II.C-2, II.C-5, II.C-6, II.D-1 and
late 1970s. In fact, the method was not widely presented to the II.D-3 all contain calculations for the Whitmore section.
engineering community until 1974, when it was discussed in Fisher
and Struik’s Guide to Design Criteria for Bolted and Riveted Joints. Areas of Possible Confusion
There are several predictable areas where confusion can arise in
These Days, It’s in the Manual dealing with a Whitmore section. We will address these individually.
An explanation of how to calculate the Whitmore section is When the effective width crosses a connected edge. Part 9
provided in Part 9 of the 14th Edition AISC Manual. A figure is of the 14th Edition AISC Manual states, “The Whitmore section
also provided in the Manual (Figure 9-1) to aid the user and is may spread across the joint between connecting elements, but cannot
shown here in Figure 3. spread beyond an unconnected edge.” All of the examples provided in
The Whitmore section is used to determine the peak tension the Design Examples where the Whitmore section spreads across the
or compression stress of an uneven stress distribution at the end joint also happen to be cases where the gusset plate edge is welded to
of the joint. It does this by establishing an effective length, which the beam flange. While it may be stating the obvious, any connection
Whitmore determined could be calculated by spreading the force that has been properly designed, such as a bolted-bolted or bolted-
from the start of the joint, 30° to each side in the connecting ele- welded double-angle or single-plate connection, can be considered a
ment along the line of force. The most common application of the connected edge when the Whitmore section passes through it.
Whitmore section is in gusset plates for bracing and truss con- When the effective width crosses a joint between 36 ksi
nections. Figure 4 shows a gusset plate that has failed in tension and 50 ksi material. There may be some confusion as to how
rupture after significant tension yielding at the Whitmore section. to use the Whitmore section for tension and compression checks
The predicted strength was in good agreement with the measured when the Whitmore section spreads across a joint between a gus-
failure load. set plate and a beam or column that have different strength levels.
One might expect that the stress distribution is uniform and that
there is no way to have two separate levels of stress. However, as
shown in Example II.C-2 in the Design Examples, we can take
advantage of the higher strength material.
Once the lower strength material (typically the gusset plate)
reaches its yield strength, it will strain and allow the load to distrib-
ute to the higher strength material (see Figure 6 on the following
page). Note that the amount of strain involved for this to occur is
negligible, as shown in Figure 5. This is an inelastic but self-limiting
deformation much like that used in the design of “simple connec-
tions,” and in this case, any tendency to rotate due to the uneven
stress distribution on the Whitmore section is limited by the sur-
rounding material that does not participate in load resistance, but
Fig. 3: Illustration of the width of the Whitmore section. would have to shear for rotation to occur. See the next question for
➤ ➤

more on this. The hybrid section may change the stress distribution
Fig. 4: Tension rupture of Whitmore section.
but use of a design stress calculated from the hybrid Whitmore sec-
The AISC Design tion will provide a gusset that performs in an acceptable fashion.
Examples that complement
the 14th Edition AISC
Manual contain numerous Fig. 5: Stress-strain curve for different yield stresses

examples of how to calcu- (Salmon and Johnson, 1996).

late the Whitmore section,
along with how it is used
in determining the tension
yielding or compression
buckling strengths of the
gusset plate. These exam-
ples are also available online
at www.aisc.org/epubs. To
Tom Murray/Virginia Tech

calculate the tension yield-

ing and compression buck-
ling strengths of a gusset
plate, where the Whitmore
section occurs over both
the gusset and beam web,


Fig. 8: Eccentricity due to gusset plate geometry.

Two options for treating an eccentric loading condition due
to gusset plate geometry include:
➤ Option 1: Adjust the gusset plate geometry to avoid eccen-
Fig. 6: Eccentric loading on Whitmore section due to differing tricity as shown in Figure 9 (a). This is the preferred solu-

material strengths. tion if it is possible to do so.

➤ Option 2: Conservatively analyze the gusset plate for a
Eccentricity when the distribution isn’t balanced. reduced Whitmore section effective width that is bal-
Another possible concern with taking advantage of the higher anced along the work line, as per Figure 9 (b). While this
strength material is eccentricity. That is, what if doing so means approach is conservative, it is a quick and easy solution.
that the resultant force is no longer centered on the work line as
in Figure 6? The effects of this eccentricity can be ignored. The
gusset plate is very rigid and the surrounding metal will restrain
any rotation that would otherwise occur on the Whitmore sec-
tion (see Figure 7).

Fig. 9: Options for dealing

with eccentricity due to

gusset plate geometry.

Another example of when geometry could be an issue is the

hanger to truss panel point connection shown in Figure 10. The
Whitmore method is not necessary for the vertical member to
Fig. 7: Gusset plate rigidity. gusset plate connection. Most of the load will be transferred to

the diagonal member through shear in the gusset plate. Cal-

Gusset plate and member web of different thicknesses. culations for a similar connection detail are shown in Design
When the Whitmore section spreads across the joint of two com- Example II.C-6 that works with the 14th Edition AISC Manual.
ponents of differing thicknesses, the distribution of force similarly This example does not include any Whitmore section calcula-
may not be uniform as for the case of materials of different strength. tions for the truss vertical member.
As explained above, the same conclusion (for the same reasons) may O ve r e s t i m a t -
be stated here in that eccentricity about the Whitmore section is ing the Whitmore
not a concern and its effects need not be calculated. section. Depend-
Gusset plates of restricted geometry. Gusset plate geom- ing on the brace
etry may have a significant impact on the tension and compres- connection config-
sion strengths available on the Whitmore section. While eccen- uration, there are
tricity as a result of differing plate thicknesses and material occasions where the
strengths is not a concern when using the Whitmore method, calculated length
eccentricity due to plate geometry may be important to con- of the Whitmore
sider because often the geometry limit also eliminates some of section could be
the stabilizing effect illustrated in Figure 7. Plate geometry may less than what one
result in a non-uniform tensile stress distribution across the
Whitmore section and no side material to restrain the rotation ➤
Fig. 10: Connection
(see Figure 8). at truss panel point.


Unbraced length for com- Note that when computing the
pression strength calculations. buckling length using an average
When calculating the buckling of the three lengths (L1, L2 and L3),
strength of a gusset plate, the L1 or L3 may be subtracted if the
approximate length of the gus- Whitmore section spreads across
set plate that will buckle must be the joint into the beam or column
determined. The two most com- as illustrated in Figure 12, Case B.
mon methods for determining Design example II.C-2 from the
this length are shown in Figure Design Examples also covers both
12. Both methods are permitted. of these methods.

Fig. 11: Overestimating the Whitmore


would initially expect. This depends on

the depth of the brace, how it is connected,
and the length of the overall connection.
Figure 11 provides an example of one case
where the Whitmore length could be cal-
culated incorrectly.

Fig. 12: Gusset plate buckling length



{ }
Whitmore doesn’t just affect gusset
The People Behind the Theory plates. While most of this discussion has
Who was Armour T. Granger? focused on Whitmore sections in gusset
Armour T. Granger came to the University of Tennessee in 1939 plates, there are other types of connections
after working for Ash, Howard-Needles & Tammen in New York where this section should be calculated.
City. According to David W. Goodpasture, professor emeritus in One example is the truss connection shown
the Civil and Environmental Engineering Department at the Uni- earlier in Figure 10. Another example is a
versity of Tennessee, Granger was “very interested in the behavior WT hanger connection, which is shown in
of joints in a Warren truss.” Figure 13 below. If a connection is being
Goodpasture said that Granger asked Whitmore and another designed to transfer axial load, an engineer
graduate student to study the joint in the early 1950s. “Whitmore will need to determine if a Whitmore sec-
made an aluminum model of the joint and . . . used electrical strain tion check is required.
gages,” Goodpasture said. “I can still remember seeing the model
in the basement of Perkins Hall. It was about five feet tall. Whitmore
wrote the engineering experiment publication based on both stu-
dents’ (M.S.) theses.”

Who was R.E. Whitmore?

According to Edwin Burdette, a professor of civil engineering at
the University of Tennessee, Whitmore was an assistant professor
in the early 1960s at the University of Tennessee where he taught a
materials course. He was a popular teacher, winning the first “Fac-
ulty Man of the Year” award given by the student chapter of ASCE
in 1964. He also went on to be a successful road builder, though he
is perhaps best remembered for his gusset plate article.

Fig. 13: WT hanger connection.

One Final Note

The 2011 AISC T.R. Higgins Lecture-
ship Award winner, Charles W. Roeder,
P.E., Ph.D., has introduced a very practi-
cal approach to the design of seismic gus-
set plates in special concentrically braced
frames. While there are no changes to
the Whitmore section, this new approach
lets the gusset plate bend line occur in an
elliptical pattern, allowing the gusset plate
connection to be more compact, which
helps reduce the size and cost of the brac-
ing frame connection. The Whitmore sec-
tion calculations for compression on these
gussets will benefit greatly from the shorter
buckling lengths. 

The authors would like to thank Tom Murray,

Charlie Carter, Tom Schlafly and Leigh Arber
for reviewing this article and for providing
valuable feedback. The quality of the article
benefited greatly from their involvement.